Why did demand fetching endure as the conventional wisdom for so long? Why are anticipatory fetch strategies receiving so much more attention today than they did decades ago?

What will be an ideal response?

The appeal of demand fetch is that it is simple (i.e., the KIS design philosophy at its
best), it is easy to implement, and it brings to main memory only those program pieces or
data pieces that are absolutely going to be used immediately.There has been a shift in the relative
value of computer resources vs. people time.Anticipatory mechanisms required considerable
overhead: Deciding which pieces to bring to main memory took significant time, it
consumed main memory resources, and anticipatory schemes could make bad decisions
(since we cannot make completely accurate predictions). In the past, this overhead was sufficiently
great to deter implementors. Today, processor speed are extremely fast compared to
disk speed, which reduces the relative overhead incurred by deciding which pieces to bring
intomain memory from disk.But today the primary is how to improve system responsiveness
without fetching pieces of programs that will not be referenced (anticipatory strategies can
still make bad decisions). Processor times is relatively inexpensive, but memory overhead
due to poor anticipatory fetching decisions can be significant.